This article considers the design process for new formative history assessments. Over the course of 3 years, my colleagues from the Stanford History Education Group and I designed, piloted, and revised dozens of "History Assessments of Thinking" (HATs). As we created HATs, we sought to gather information about their cognitive validity, the relationship between the constructs targeted by the assessments, and the cognitive processes students used to answer them. Three case studies trace the development of different HATs through analyses of draft assessments, student responses, and think-aloud protocols. Design principles specific to formative history assessments emerged from these analyses: (1) assessments must be historically accurate, (2) assessments must target specific historical constructs, (3) assessment structure must align with targeted constructs, (4) assessments must yield useful information for teachers, and (5) pilot data are indispensable for refining HATs. These finding suggest the need for increased attention on the construction and validation of new assessment materials for the history classroom.